McDonald, Dan wrote:
>> From: clamav-users-bounces@lists.clamav.net on behalf of Luca Gibelli
>> Hello jacusy,
>
>>> this night my clamd-process terminated with an error. The reason was
>>> that freshclam took too long to do its update, so that clamd could not
>>> lock the database. So clamd exited. But this behaviour is very fatal
>>> because the mail system (postfix with amavis) relys on clamd, so if it
>>> is down, the whole mail traffic is blocked!! Caused of an error while
>>> updating......
>> amavisd-new falls back to clamscan by default, if clamd is not available.
>
> Right, but my system is sized for operation with clamd. When clamd stops, the memory utilization goes through the roof and the machine basically stops because it is spending 80-90% of the cpu time in IOWAIT, trying to swap out enough ram to load the whole clamav database for every message...
> And mail backs up... At least it doesn't drop any mail.
>
And none of this discusses the real issue that clamd should not have
been unavailable in the first place. I understand that daemons fail and
critical ones should have monitoring in place. This is a no brainer.
However, it is illogical that clamd would die completely due to issues
with a recently downloaded definition file. Why can it not just roll
back to the old, previously working, definitions? Can someone please
explain this? Im having trouble trying to comprehend the current behavior.
_______________________________________________
Help us build a comprehensive ClamAV guide: visit http://wiki.clamav.net
http://lurker.clamav.net/list/clamav-users.html
(text/plain)